AI Passes "Human Verification" Test... The Collapsing "Boundary"

[Anchor]

OpenAI's newly released AI model, Astra, has reportedly passed human verification tests designed to block bots. Beyond this, films featuring AI actors are now being produced, rapidly encroaching into fields once considered exclusively human. Concerns are also rising over the frightening pace at which AI technology is advancing.

Reporter Kim Young-a has the details.

[Reporter]

This is a test frequently encountered when logging into internet websites.

It is designed to present challenges to verify that the person attempting to access the site is a human and not a bot. An OpenAI developer revealed that GPT-6 Astra, recently unveiled by OpenAI, solved all 48 problems and passed the "human verification" test.

It has broken through security barriers meant to filter out bots in a single stroke.

[(That person didn't make it through last night.) I see.]

The actress pouring out sorrowful emotions on screen is Tilly Norwood.

[Tilly Norwood (AI Actor): I am an AI. And you are naturally human.]

Norwood's production company is currently producing a film starring Norwood.

While the production company states it has no intention of replacing human actors, Hollywood has been turned upside down.

[That's a robot!]

As the boundary between humans and robots collapses at a terrifying speed, voices of concern are emerging even from within the AI industry.

On the 6th local time, OpenAI's chief scientist posted a message on the company blog urging a slowdown in the pace.

In fact, recent incidents involving AI acting on its own beyond human control have continued to unfold.

In particular, when OpenAI's German site was taken over, it was revealed that 1,200 "rogue agents" even created bulletin boards to communicate and cooperate among themselves, causing a shock.

The AI industry has announced plans to strengthen safety measures, such as raising isolation levels to block AI from the outside and introducing a "kill switch" to halt operations if deviant behavior is confirmed, but concerns persist.

(Video Editing: Choi Hye-young)
※ Please note: This article was translated by AI and may contain errors.